To investigate the use
of Cepstral measures like Cepstral Peak Prominence (CPP) and smoothened
Cepstral Peak Prominence (sCPP) to assess dysphonia severity in individuals
with Gastroesophageal reflux disease (GERD).
study included 2 groups. The first group with consisted of 20 patients,
consisting of 11 males and 9 females aged between 20 to 60 years diagnosed with
Gastroesophageal reflux disease (GERD) whose voice sample were selected from
the CSL4500 (KayPentax) database. The second group consisted of normal age and
gender matched individuals.
Subject selection criteria:
Those with a clinical
diagnosis of Gastroesophageal reflux disease (GERD) on stroboscopic examination
Those who did not have
hearing loss, hormonal problems, respiratory illness, psychological problems,
neurological problems, smoking and alcohol consumption.
Those who reported of
Those who reposted of
have hearing loss, hormonal problems, respiratory illness, psychological
problems, neurological problems.
recordings were carried out in a sound treated lab following ANSI S12.2-2008 standards.
Profile (MDVP) modules of computerized Speech Lab (CSL) software; Model 4500
A Shure mic belonging
Speech tool program Version 1.65 by James Hillenbrand
Stroboscope Direct Rigid Laryngoscope (Xion Endostrob DX) was used to
visualize the vocal folds
Detailed case history which included onset of problem,
duration of symptoms, associated voice problems and medical management was
taken for individuals with GERD. Additional information regarding the
individual’s lifestyle like smoking or drinking habits, daily use of voice,
diet, history of hearing loss, hormonal problems, respiratory illness,
psychological problems, neurological problems was also noted for both groups
Instructions and recording:
had been made in a sound-treated lab on CSL4500 using a Shure microphone belonging
to CSL, KayPentax. All participants had been instructed to vocalize a sustained
vowel /a/ for more than 8 seconds with a voice reflecting their usual pitch,
strength, and quality in a sitting position. All the files were saved as .wave
files on the CSL4500 system. These files were then transferred to a laptop (DELL,
Inspiron 15) for analyses using the Speech tool program Version 1.65
samples recorded from both these groups was edited to retain the middle and
stable portion of vowel /a/ for duration of 3 sec. Sustained vowel samples were
analyzed by measuring CPP and sCPP, using the algorithm and computer software
program (Speech tool program Version 1.65)
for calculating sCPP obtained from Hillenbrand. The computer algorithm then
calculated CPP and sCPP for each sample. The CPP, in CSL was calculated by
subtracting the value of the peak from the apparent baseline signal.